With that in mind I decided to enter the contest. First prize is 5K and who couldn’t use some extra cash, right? Although I found out about the contest a little late in the game, I was reasonably sure that I could put something together that, even if I didn’t win or place at all, would still be useful to someone. However, putting together an app in 30 days with all kinds of other things going on isn’t exactly a non-complex task. I also have a job and a family and both of those place other requirements on my time. So, I had to utilize my time effectively and come up with a decent idea that wasn’t too grandiose.

I thought a nice compliment to the contest submission would be a post that essentially outlined my development and the decisions ( and mistakes ) I made. It was certainly a bit of a learning process. In my research I came across something that, in my own mind, changed the course of things and hopefully made my efforts more focused.

Sifting through the data:

The first thing of course, and since it was a requirement of the contest, that had to be done was sift through the mountains of available datasets. The whole time you’re doing it the glaring question is, “What can you possibly do with this data?” Part of the requirement for this app is that you must use at least one set of data from DOC. You are more than welcome to use any other data, but at a minimum one must be from DOC. This was a little confusing at first – trying to determine what is from where – but I think I ultimately cleared that up for myself.

One caveat when starting to look at data available is that it might be wise that you either be knowledgable about the data in the first place or know someone who is. Some of the data has good documentation and some does not. I found this out when I started working on my first idea.

The first idea:

At first I started looking at Health Care data. I had a good idea of what might be possible, but I wasn’t sure I was using DOC data. Also, when I looked at the health care datasets available and the data that lived within, I was able to use it and possibly form the necessary relations between the datasets. However, at times I questioned whether these relationships were correct. Ultimately I decided I needed to look for a different idea that was smaller in size.

The second idea:

Consumed with the strong belief that my first idea, no matter how feature-rich and packed with volumes of data it was, wasn’t entirely going to work, I set out to find a new idea with only three weeks until the submission deadline. I decided to take a look at the Farmers’ Markets dataset. The initial set I found only had the names and locations of markets across the USA. I built a quick mapping app to try out the data and it seemed to work, but there was one problem . . . the latitude and longitude were reversed in the data. After I fixed that all seemed pretty good.

This idea, in my opinion, was probably going to work for me. I really only had three weekends to spend on the idea and even those were partially spoken for. At work my team is really focused on our Best of the Web submission this year so I can’t miss work and there is the likelihood I may have to spend a little weekend time at work. I did a little Googling for Farmers’ Market apps and there was no over-abundance of those. One iPhone app I did find seemed to have a few disappointed customers. The clincher was the fact that I found some more data that had more information packed in it.

With all this in mind I decided to determine exactly how I was going to build the app.

The requirements:

The rules state that you can build any kind of app – be it an iPhone, iPad, Android, web browser or even desktop app. I didn’t want to focus on any specific platform. If I was going to develop this app it was going to be available for every platform – although the desktop is the one exclusion to that. On the desktop the user is more than welcome to use a web browser. In my own mind, unless there is some compelling reason or functionality requirement that specifically targets a device, an app should be built such that any Internet-enabled device can reliably access it.

For the purposes of this app I decided the following:

The user will access the URL of the Farmers’ Market Finder app using either an Android, IPhone, iPad device or a web browser. Regardless of the fact that they have set their geo-location functionality either on or off they will be able to provide their location. While allowing their location to be determined by geo-location functionality will be more accurate, they can alternatively provide their city and state. Farmers’ Markets in their vicinity will be revealed using Google Maps. The will be able to click or tap on a map icon and the name of the market and Street View will be shown in the Info Window. They may click or tap on a button inside the info window to get more information.

If they have elected to get more information they are taken to a new URL in which they may view the name, address and URL of the website for the market. They may also view the offerings available at that market and lastly, they may view driving directions to the market.

Building it:

Obviously Google Maps plays a large part of the functionality behind the app. However, what I considered to be the biggest win was Bootstrap from Twitter. With it I was able to achieve the cross platform goals I originally decided in the requirements. The baked-in responsive design of Bootstrap allows for any device to work with this application.

The backend is Django with a relational database. Last, but not least, it is hosted by an awesome web host – Webfaction.

For the most part, development on this project was grabbed whenever possible. Small things cropped up from time to time in testing. For the most part all browsers work fantastic. Firefox does something weird that I haven’t yet pinpointed. Perhaps the geo-location backend for it is just slow, but it seems to take longer than I like to get the geo-data back. Also, sometimes it works unreliably. It’s rare, but it sometimes happens.

Also, this:

  navigator.geolocation.getCurrentPosition( initiate_position,handle_errors );

had to be changed to this:

navigator.geolocation.getCurrentPosition(function(position) {
 clearTimeout(location_timeout);
 initiate_position(position);
  }, function(error) {
   clearTimeout(location_timeout);
    handle_errors(error);
 });

In order for things to work in Firefox.

As I said, I’m not 100% sure why, but I didn’t want to leave Firefox out of potential compatible browsers.

But does it work?

I believe so. You can take a look at the screen shots here, but for the best experience try it out for yourself:

Try the Farmers’ Market Finder!

Please, if you give it a try leave some comments and let me know what you think. Does it work for you? Have any problems. Let me know. The app has yet to be accepted as an “official” entry, but it has been submitted.

Open Data App Development

About a month or so ago I found an announcement for an app development contest. I’ve never done that so I thought I might give it a shot. The grand prize is 5K. The catch is that you have to utilize data from data.gov. There are other requirements. I’ll touch on those in another post because the entry deadline is May 1 – so I’ll probably post again with more details at that time.

Mostly I just wanted to post today for a couple of reasons:

1. I really need to get some content out here. I’ve been making all kinds of changes to stuff on this site, but nothing really satisfies me to the point that I want to make it stick.

2. The main reason for number 1 is that I’ve come to a conclusion about things and they need to get solidified. I’m working on that with what little free time I can grab.

There’s lots and lots of data at data.gov. If you’re into developing apps you know that your app is nothing without data. The challenge is finding data and then trying to do something fairly creative with it. Perhaps you have a creative idea and the only thing left is to get the data, but for me it was the other way around. There was lots of data but no exceptionally creative idea behind it.

The other challenge is dealing with the data. For the most part, there is what there is. Along the way I looked at several datasets only to find that there were fields that would have been nice, but didn’t exist. In some cases there were complimentary datasets, but no indication of how they fit together.

Some data may be incorrect. I found data that was obviously incorrect. Those problems will just have to exist unless I can find the time to fix it. In the mean time the best I can do is display a disclaimer.

A few details. I built the app with the following:

1. Django
2. Bootstrap from Twitter
3. Data from Data.gov
4. Google Maps
5. Awesome web hosting from Webfaction

After a brief hiatus . . . .

I’m not sure how long it’s been since I started up this site. It’s probably getting pretty close to a year. Lately, however, I decided to close out a bunch of old stuff and try something new.

Right now I’m experimenting with this Roots theme that utilizes Twitter’s Bootstrap CSS framework.

While the site is predominantly empty right now, I’m re-building and will hopefully have something worthwhile out here before long.